Added unbounded array patterns

[K-Phoen]

Can be useful when you want to test the structure of an array without knowing its size.

[norberttech]

Hi @K-Phoen feature looks cool. It would be nice if you could you provide some real use case for it, you know maybe some kind of scenario.
I'm just not sure about the @...@ pattern. I think its not enought descriptive and might be a bit confusing. Any other suggestions?

[K-Phoen]

I'm writing (behat) tests for an API and this feature allows me to "strictly" check the first N results that are returned and just check the structure of the rest.
I'd like to write things like this in my features:

Then the response should match json:
'''
{
    "results": [
        {
            "id": @string@,
            "foo": "bar",
            "bar": 42
        },
        {
            "id": @string@,
            "foo": @string@,
            "bar": @integer@
        },
        @...@
    ]
}
'''

Concerning the pattern, I haven't a lot of other ideas. Would you prefer @rest@ instead of @...@?

[norberttech]

Oh yest, it might be useful for api testing. But I'm still not convinced about the pattern. I think the name should point out that this pattern can be used only for arrays. Maybe @array_wildcard@, or @array_rest@? @defrag what do you think about it?

[defrag]

Hmm well i guess it makes sense for some scenarios, but i have a feeling with this implementation its confusing how to use it. Maybe we should introduce some repeatable array pattern?

[K-Phoen]

Can you be more explicit?

[defrag]

Im thinking something like:

Then the response should match json:
'''
{
    "results": [       
        {
            "id": @string@,
            "foo": @string@,
            "bar": @integer@
        }
    ].repeat(0...1|Inf)
}
'''

or something similar. I honestly dont see the point of matching two first values and then the rest pattern.

[norberttech]

Looks like @defrag solution solves @K-Phoen problem, right? To be honest I like it more, its more desriptive.

[defrag]

But i guess this would be much more work to achieve that :P

[K-Phoen]

Yep, I also prefer this solution :)
I don't know when I'll be able to work on an implementation though.

[jakzal]

I honestly dont see the point of matching two first values and then the rest pattern.

Here's a case: verify if a new entry was added to the response and ignore the rest.

[defrag]

@jakzal but in scenario described here we can verify only if its on one of the first position. We cant check if it was addes as last one.

[norberttech]

To solve @jakzal use case I can think about two possible solutions

Then the response should match json pattern:
'''
{
    "results": @[email protected](["id": "@integer@", "name": "New"])
}
'''

Then the response should match json pattern:
'''
{
    "results": [
      @...@,
      {
        "id": @integer@
        "name": "Norbert" 
      },
      @...@
    ]
}
'''

[norberttech]

btw. @K-Phoen I belive you can achieve your goal with Expression expr() matcher.

[norberttech]

@K-Phoen could you please rebase this PR against master? It should be enough before 2.0 release where we plan to add pattern expanders.

[K-Phoen]

Done :)

[norberttech]

Thanks!